ARTICLE IN PRESS

Available at www.sciencedirect.com

WAT E R R E S E A R C H 4 0 ( 2 0 0 6 ) 2 3 8 7 – 2 3 9 6

0043-1354/$ - see frodoi:10.1016/j.watres

�Corresponding aE-mail address: rg

journal homepage: www.elsevier.com/locate/watres

Quantitation of hepatitis A virus and enterovirus levels inthe lagoon canals and Lido beach of Venice, Italy, usingreal-time RT-PCR

Michael A. Rosea, Arun K. Dharb, Hilary A. Brooksa,Fulvio Zecchinic, Richard M. Gersberga,�

aGraduate School of Public Health, San Diego State University, San Diego, CA 92182, USAbDepartment of Biology, San Diego State University, San Diego, CA 92182, USAcLaboratorio di Analisi di Microbiologia Ambientale (LAMA), Consorzio Interuniversitario Nazionale ‘‘La Chimica per l’Ambiente’’ (INCA),

Via delle Industrie, 21/8, I-30175 Marghera, Venezia, Italy

a r t i c l e i n f o

Article history:

Received 7 September 2005

Received in revised form

24 March 2006

Accepted 29 March 2006

Available online 5 June 2006

Keywords:

Hepatitis a virus

Enteroviruses

Real-time RT-PCR

Risk assessment

Venice Lagoon

nt matter & 2006 Elsevie.2006.03.030

uthor. Tel.: +1 619 594 290ersber@mail.sdsu.edu (R

A B S T R A C T

In order to assess the microbial water quality of the lagoon canals of Venice, Italy and

nearby beach on Lido island, a study was conducted using real-time RT-PCR to enumerate

levels of hepatitis A virus (HAV) and enteroviruses in these marine waters over a 3-year

period from 2003 to 2005. A total of 17 sites (9 lagoon canal and 8 beach sites) were assayed.

For the canals of the Venice Lagoon, 78% were positive for both HAV and enteroviruses, with

levels ranging from 75 to 730 and 3 to 1614 genome copies/L, respectively. At Lido beach,

HAV was never detected, but enteroviruses were detected in all Lido beach samples at levels

ranging from 2 to 71 genome copies/L. There was a statistically significant correlation

between thermotolerant coliform densities and HAV levels (p ¼ 0.0002), but the relationship

between thermotolerant coliform densities and enterovirus levels was not significant

(p40.05). Despite the fact that enteroviruses were detected at low levels in the surfzone at

Lido beach, the risk for enteroviral infection (calculated using the beta-Poisson model) for

recreational exposure from swimming, was in the range of 1.9� 10�3–6.1� 10�2, yielding a

disease risk at or below the level (5% for gastroenteritis) deemed acceptable by European

Guide standards.

& 2006 Elsevier Ltd. All rights reserved.

1. Introduction

The unique location of the city of Venice, Italy, built on a

number of islands in the middle of the Venice Lagoon, has

rendered it impossible to construct a sewage treatment

infrastructure for the city and surrounding islands (Pavoni

et al., 1990). The lagoon receives the untreated sewage from

Venice with an organic and pathogen loading equivalent to

more than 400,000 persons during the tourist season (Orlob

et al., 1991). Although actual bathing is prohibited in the

r Ltd. All rights reserved.

5; fax: +1 619 594 6112..M. Gersberg).

lagoon itself, recreational beaches exist on the littoral strip of

the Adriatic sea (e.g. Lido beach about 3 km west of the

northern Lagoon opening) which may be adversely impacted

by contaminated lagoon waters flowing out on the outgoing

tide. Additionally, Venice Lagoon is unusual insofar as there is

potential to impact human health through combined expo-

sure routes, some of which are not typical (Johnston et al.,

1993). For example, transport in Venice is predominantly by

boat. Disturbance of the contaminated water by motorized

vessels may create aerosols which may pose an inhalation

ARTICLE IN PRESS

WAT E R R E S E A R C H 4 0 ( 2 0 0 6 ) 2 3 8 7 – 2 3 9 62388

health hazard (Blanchard, 1989). Moreover, due to sea-level

rise combined with soil subsidence, flooding events have

become more frequent, occurring some 50 times a year in the

past decade (Bernstein and Cecconi, 1996). Such floods which

inundate streets and businesses increase the number of

waterborne disease exposure scenarios. A project with an

estimated cost of 4.5 billion US $ has recently begun to

construct movable tidal barriers to stop flooding of the city of

Venice (the MOSE Project). However, a number of environ-

mental concerns about this barrier project remain, particu-

larly that the operation of these barriers could decrease the

tidal flushing of the Lagoon and exacerbate an already serious

sewage pollution problem.

There is currently a lack of microbiological and epidemio-

logical studies that provide an assessment of the risk of

human disease associated with sewage disposal into the

Venice Lagoon (Aimo et al., 1999). Vazzoler and Stradella

(1999) detected enteroviruses (by tissue culture techniques) in

the canals of the city of Venice, but did not present

quantitative results for virus levels that could be used in a

risk assessment for bathing at Lido beaches or for other

scenarios involving exposure to waters of the Venice Lagoon.

The present study determined levels of both hepatitis A virus

(HAV) and enteroviruses in the lagoon canals of Venice and

the surfzone at nearby Lido beach by real-time RT-PCR, in

order to better assess human disease risk from recreational

exposure to these marine waters.

2. Methods

2.1. Sampling sites

A total of 17 water samples (Fig. 1 and Table 3) were collected,

including seven samples from the Grand canal (at Rialto

bridge), one sample from each of two interior canals, the Rio

di San Marcuola and the Rio di Fuseri, and eight samples from

a recreation beach at Lido (near Via Santa Maria Elisabetta

(SME)), on the littoral strip of the Adriatic Sea beach about

3 km west of the northern Lagoon opening (Fig. 1 and Table 3).

Fig. 1 – (A) Map of Venice Lagoon showing location of Lido beach

sampling sites.

Samples ranged from 2 to 12 L and were collected over three

consecutive years in the summers of 2003, 2004, and 2005

(Table 3).

2.2. Virus concentration

Each sample was processed within 1–2 h of collection follow-

ing a published protocol by Katayama et al. (2002). Seawater

samples were filtered at a constant rate via a vacuum pump

through a series of Whatman filters (of 11 and 2.5 mm pore

size) to reduce particulate matter. Although it is well under-

stood that viruses can adsorb to particles in the environment,

removal of particulates is necessary for PCR assays. Samples

were then applied to a type HA 0.45-mm negatively charged

membrane (Millipore, Burlington, MA, USA). The negatively

charged filter was washed with 200 mL of 0.5 mM H2SO4 to

remove cations, and the virus was eluted from the filter with

10 mL of 1 mM NaOH, into a tube containing 0.1 mL of 50 mM

H2SO4 and 0.1 mL of 100�TE buffer (Sigma-Aldrich, St. Louis,

MO, USA). The filtrate was then concentrated to 450mL volume

by centrifuging the samples in a Centriprep Concentrator

(YM-30, Millipore) at 1500g for 15, 10, and 5 min. Total RNA

was extracted from the 450mL filtrate using TRI ReagentTM

(Molecular Research Center Inc., Cincinnati, OH, USA) and the

RNA pellet was dissolved in 40mL of TE buffer (pH 8.0).

2.3. Quantitation of HAV by SYBR Green real-time RT-PCR

Procedures for cDNA synthesis and SYBR Green real-time RT-

PCR were performed as described by Brooks et al. (2005),

except a BioRad iCycler real-time thermocycler was used

instead of the Applied Biosystems GeneAmp 5700 Sequence

Detection System for real-time RT-PCR. First strand cDNA

(40mL) was synthesized from 19.5mL of RNA using random

hexamers. Sample cDNAs were diluted 1:10 and 1:100 with

DNase, RNase-free water containing sonicated herring sperm

DNA (5 ng/mL) as carrier DNA (Leutenegger et al., 1999). The

SYBR Green RT-PCR amplification was carried out in a 25mL

reaction volume that contained 7.1mL of 2� SYBR Green

sampling site. (B) Map of Venice canals showing location of

ARTICLE IN PRESS

Table 1 – List of primers and probe used for HAV and enterovirus assays by conventional and real-time RT-PCR

RT-PCR Primer name Primer sequence (50–30) (%) GC Ampliconsize (bp)

References

Conventional

HAV HEPA1 Forward: GTT TTG CTC CTC TTT

ATC ATG CTA TG

39 247 Brooks et al.

(2005)

HEPA2 Reverse: GGA AAT GTC TCA GGT

ACT TTC TTT G

40

Real-time

HAV HAV1FWD Forward: TAC AGA GCA GAA TGT

TCC TGA TCC

46 76 Brooks et al.

(2005)

HAV3RVS Reverse: TCC CCT ATT GGC TTT

CCC TT

50

Enterovirus EV1FWD Forward: GGC CCC TGA ATG CGG

CTA AT

40 151 MGB AlertTM

Real-Time

PCRKit

(Xanogen)

EV1RVS Reverse: CAA TTG TCA CCA TAA

GCA GCC A

55

Probe MGB-EDQ-CTT TGG GTG TCC

GTG T-Q14-FAMa

44

a MGB ¼minor groove binder, EDQ ¼ eclipse dark quencher, FAM ¼ 6-carboxy fluorescein.

WAT E R R E S E A R C H 40 (2006) 2387– 2396 2389

Master Mix (Applied Biosystems, Foster City, CA, USA), 300 nM

each of the forward and reverse primers (Table 1), and 1 mL of

undiluted stock or diluted cDNA. Each sample had three

replicates to ensure the reproducibility of results. The thermal

profile for SYBR Green real-time RT-PCR was 95 1C for 10 min,

followed by 40 cycles of 95 1C for 10 s and 60 1C for 1 min.

2.4. Quantitation of enterovirus by molecular beacon real-time RT-PCR

Real-time PCR was accomplished using a One Step RT-PCR Kit

(Qiagen, Valencia, CA, USA) and an MGB AlertTM Enterovirus

Real-Time PCR Kit (Nanogen, San Diego, CA, USA). The

enterovirus kit contained a 20� primer mix as well as a

20� MGB Eclipse Probe (Table 1) directed toward the 50

untranscribed region (UTR) of enteroviruses (coxsackie A and

B, echoviruses, polioviruses, and enteroviruses 68-71). The

RNA samples were diluted 1:10 and 1:100 with DNase, RNase-

free water containing sonicated herring sperm DNA (5 ng/mL)

as carrier DNA (Leutenegger et al., 1999). Each 50mL reaction

mixture contained 17mL of RNAse-free water, 10mL of 5�

Buffer, 10mL of 5� Q-Solution, 2.0 mL of dNTP Mix, 2.0 mL of

Enzyme Mix (all components of the Qiagen kit), 2.5mL of 20�

forward/reverse primer mix (Table 1), 2.5 mL of 20� MGB

Eclipse Probe (Table 1), and 4.0 mL of undiluted or diluted

template RNA.

Samples were run in duplicate on a BioRad iCycler real-time

PCR system. The real-time PCR conditions were as follows:

reverse transcription for 30 min at 50 1C, polymerase activa-

tion for 15 min at 95 1C, 50 cycles of denaturation for 10 s at

95 1C followed by annealing/detection for 30 s at 56 1C and

extension for 30 s at 76 1C, and a final extension step for

10 min at 76 1C.

2.5. Cloning and sequencing of HAV and enteroviruscDNA

Samples found positive for HAV and enterovirus by real-time

RT-PCR were cloned and sequenced. A 247 bp HAV cDNA was

amplified by conventional RT-PCR following a published proto-

col (Brooks et al., 2005). The primers for HAV amplification are

given in Table 1. Amplified cDNAs were separated by electro-

phoresis in a 2% agarose gel and eluted from the gel using a

Qiagen gel extraction kit (Qiagen, Inc., Valencia, CA, USA). In

order to clone the enterovirus cDNA, real-time RT-PCR ampli-

fied cDNAs were run in a 2% agarose gel, and gel-purified using

a Qiagen QIAquick gel extraction kit. The enterovirus and HAV

gel-purified cDNAs were cloned into a TOPO cloning vector

(Invitrogen, Carlsbad, CA, USA). Plasmid DNA was isolated from

recombinant clones and three to five clones were sequenced for

each sample using the vector-derived T7 primer.

2.6. Sequence alignment and phylogenetic analysis

Nucleotide sequences of HAV and enterovirus clones were

BLAST searched and identified based on similarity to Gen-

Bank database entries. Multiple alignments and phylogenetic

analyses were performed using MEGA version 3.0 by Kumar et

al. (2004). Kimura’s two-parameter distance was calculated

using transitions and transversions and a neighbor-joining

tree was built. The confidence of reconstructed clusters was

tested by bootstrapping with 1000 replicates.

2.7. Generation of HAV and enterovirus standard curvesby real-time RT-PCR

An HAV standard curve was generated using plasmid DNA

(8�107 copies/mL) with inserted cDNA from HAV strain HM

ARTICLE IN PRESS

WAT E R R E S E A R C H 4 0 ( 2 0 0 6 ) 2 3 8 7 – 2 3 9 62390

175 (VR-2089; ATCC, Manassas, VA, USA). Serial dilutions

(from 8 to 8�104 copies/mL) were prepared in TE buffer

(Sigma-Aldrich). Plasmid containing enterovirus cDNA

(1�107 copies/mL) was obtained from Nanogen and a dilution

series (101–104 copies/mL) was prepared in TE buffer. Real-time

RT-PCR was performed in triplicate for each dilution of HAV

and enterovirus plasmids. Standard curves were created by

plotting the log of the number of HAV and enterovirus

genome copies versus their corresponding CT values and

creating a best-fit line through these points (Fig. 2). The cycle

threshold (CT) is defined as the PCR cycle at which an increase

in the fluorescence above the baseline signal is first detected.

The CT value is inversely related to the genome copy number.

Using the standard curves, HAV and enterovirus levels in the

canal and beach samples were calculated with the following

equations. Concentrations were calculated assuming that no

viral genomes were lost during the synthesis of cDNA

(Haramoto et al., 2005; Deffernez et al., 2004; Mohamed et

al., 2004),

fHAV genome copies=Lg

¼fð1� 10½ðCT�36:8Þ=�3:6�

Þðdilution factorÞð40Þð40=19:5ÞgfLiters of seawaterg

, ð1Þ

fEnterovirus genome copies=Lg

¼fð1� 10½ðCT�38:2Þ=�3:4�

Þðdilution factorÞð10ÞgfLiters of seawaterg

. ð2Þ

Fig. 2 – Standard curves of the real-time RT-PCR assays: (A) detec

8�103 (m), 8�102 (’), 8�101 (.), 8�100 (E) genome copies. (B)

at 104 (m), 103 (’), 102 (.), 101 (E) genome copies. HAV (C) and en

CT value versus the log of the number of viral genome copies. E

measurements.

2.8. Calculation of HAV and enterovirus recoveryefficiencies

Two 1 L seawater samples were seeded with known titers of

virus, one with poliovirus 2 (VR-301, W-2 strain; ATCC,

Manassas, VA, USA) and the other with HAV (VR-2089, Strain

HM 175, clone 1; ATCC) prior to filtration (Table 2). The same

amount of each virus was also spiked directly into a paired

concentrated seawater sample, following filtration, but before

RNA extraction. Real-time PCR was performed and copy

numbers were determined using the standard curves (Table

2). The recovery assay was performed twice for each virus and

the HAV and enterovirus recoveries were calculated by

dividing the number of virus genome copies in the filtered

samples by the number of copies in the unfiltered samples

(Table 2).

2.9. Detection of thermotolerant coliforms in Venice canaland Lido beach samples

To determine the thermotolerant coliform levels, 100 mL

water samples were collected and processed within 2 h of

collection. The membrane filter procedure (MF) was used to

enumerate thermotolerant coliform concentrations (Ameri-

can Public Health Association (APHA), 1992). Up to three 10-

fold serial dilutions of each water sample were applied to

cellulose acetate filters and coliforms were grown on M-FC

media for 24 h at 44.5 1C. Thermotolerant coliform colonies

tion of serial dilutions of HAV inserted plasmid at 8�104 (K),

Detection of serial dilutions of enterovirus inserted plasmid

terovirus (D) standard curves were generated by plotting the

rror bars indicate the standard error of triplicate

ARTICLE IN PRESS

WAT E R R E S E A R C H 40 (2006) 2387– 2396 2391

were counted and densities then calculated without con-

firmation. The detection limit for this method is 1 CFP/100 mL.

3. Results

3.1. Recovery of spiked HAV and poliovirus from seawater

In order to determine the efficiency of our virus extraction

and concentration protocol, seawater samples were seeded

with known amounts of HAV or poliovirus on two occasions

and virus levels were quantified using the real-time standard

curves (Fig. 2). The mean percent recovery was 12% for HAV

and 71% for poliovirus (Table 2).

3.2. Detection and quantitation of HAV and enteroviruslevels in Venice canal and Lido beach samples

Real-time RT-PCR was performed to detect and quantitate

levels of HAV and enterovirus in 17 samples from the Venice

Table 2 – Recovery efficiencies of HAV and poliovirus 2(seeded into 1 liter of natural seawater) by negativelycharged membrane followed by centrifugal ultrafiltration

Virus Spikedvirus

(genomes)

Recoveredvirus

(genomes)

Meanrecovery

(%)

HAV 12,568–32,271 1760–3240 12

poliovirus 2 591–775 451–516 71

Table 3 – Levels of HAV and enterovirus as determined by realdensities, in Venice Lagoon canals and the beach at Lido

SampleID

Location Date Samplesize (L) err

GC1 Grand canal (Rialto) 5/12/2003 3.0

GC2 Grand canal (Rialto) 5/13/2003 3.0

GC3 Grand canal (Rialto) 5/14/2003 3.0

GC4 Grand canal (Rialto) 5/15/2003 3.0

GC5 Grand canal (Rialto) 5/25/2004 4.0

GC6 Grand canal (Rialto) 5/27/2004 4.0

GC7 Grand canal (Rialto) 5/31/2004 2.0

IC1 Interior canal

(Marcuola)

5/27/2004 2.0

IC2 Interior canal

(Fuseri)

6/1/2004 2.0

LI Lido beach 5/26/2004 8.0

L2 Lido beach 5/31/2004 8.0

L3 Lido beach 6/1/2004 4.0

L4 Lido beach 5/24/2005 11.5

L5 Lido beach 5/25/2005 10.5

L6 Lido beach 5/26/2005 9.0

L7 Lido beach 5/27/2005 7.5

L8 Lido beach 5/30/2005 6.0

a ND ¼ Non-detectable.

Lagoon canals and Lido beach (Table 3). Samples that were

positive for either virus using real-time RT-PCR, were further

confirmed by sequencing and were then quantitated using

the standard curves (Fig. 2). Concentrations were calculated

assuming that no viral genomes were lost during the

synthesis of cDNA (Haramoto et al., 2005; Deffernez et al.,

2004; Mohamed et al., 2004), For each sample, the value from

the dilution which exhibited the highest number of

genome copies (i.e. showed the least inhibition) was used in

Table 3.

HAV was successfully detected in five of seven Grand canal

samples, and in both interior canal samples, at levels

(uncorrected for recovery efficiency) ranging from 75 to

730 genome copies/L, but was never detected in samples from

the beach at Lido (Table 3). Enterovirus was also successfully

detected in five of seven Grand canal samples, and in both of

the interior canal samples. However, unlike HAV, enterovirus

was also detected in all eight Lido beach water samples (Table

3). Enterovirus levels (uncorrected for recovery efficiency)

ranged from 3 to 1614 genome copies/L in the Venice Lagoon

canal samples and from 2 to 71 genome copies/L in the

seawater samples from Lido beach (Table 3).

The lowest viral concentrations we detected in our sea-

water samples via real-time RT-PCR and confirmed by

sequencing were 1.9 genome copies of HAV and 1.2 genome

copies of enterovirus per PCR reaction. Due to differences in

the volumes of water collected and the amounts of RNA used

in the PCR reactions, these numbers correspond to lowest

detection limits of 13.6–77.9 genomes/L for HAV and

1.2–7.0 genomes/L for enterovirus.

-time RT-PCR, as well as thermotolerant coliform bacterial

Thermotolant coliforms

(CFP/L)

HAVconcentration(genomes/L)

Enterovirusconcentration(genomes/L)

40,000 NDa 3

30,000 75 ND

8000 270 4

27,000 ND ND

5000 128 1614

87,220 94 234

91,670 108 35

540,000 730 164

9110 128 51

7 ND 71

60 ND 19

ND ND 15

155 ND 2

26 ND 2

7 ND 2

24 ND 12

13 ND 3

ARTICLE IN PRESS

Table 4 – Enteroviral types and the designations used to create the neighbor-joining tree in Fig. 3

Viral type Designation Genbank accession number

Human echovirus 13 isolate BE00-51 50 UTR, partial sequence Echovirusl3 AF521464

Human enterovirus 90 genomic RNA, complete genome Enterovirus90 AB192877

Human enterovirus B strain EV30_18733_02 50 untranslated region EnterovirusB AY271469

Human poliovirus 2 genomic RNA, complete sequence Poliovirus2 POL2CG1

Human echovirus 11 strain Pz 87 50 non-translated region, partial sequence Echovirusll AF447476

Human poliovirus 1 isolate CHN-Jiangxi 89-1, complete genome Poliovirusl AF111984

Coxsackievirus A16 G-10, complete genome CoxsackievirusA16 CAU05876

Fig. 3 – A neighbor-joining tree derived from an alignment of

151 base pairs of the 50 UTR from enteroviral standards and

our specimens from Venice, Italy. Kimura’s two-parameter

distance was calculated using transitions and transversions

and the confidence of reconstructed clusters was tested by

bootstrapping with 1000 replicates.

WAT E R R E S E A R C H 4 0 ( 2 0 0 6 ) 2 3 8 7 – 2 3 9 62392

3.3. Cloning and sequencing of HAV and enteroviruscDNA from Venice Lagoon and Lido beach water samples

Successful HAV amplification was obtained for 7 out of 17

samples. Multiple alignments of the HAV sequences showed

100% similarity among the isolates. A BLAST (Altschul et al.,

1997) search showed 100% similarity with the VP1–VP3 gene

of HAV strains (accession #AY441443). Of the 17 samples, 15

provided successful enterovirus amplification by real-time

RT-PCR. The length of the amplicon was 151 bases. Three to

five clones were sequenced for each sample with a total of 60

clones. A BLAST search using the 151 nucleotide sequence

showed that all the clones had a similarity to the 50-

untranslated region (UTR) of enteroviruses in the database

entries. Seven types of enteroviruses were identified among

the clones sequenced (Table 4). A neighbor-joining tree

constructed from an alignment of the 151 bases nucleotide

sequence of 50-UTR revealed two major clusters (Fig. 3). The

larger clade contained poliovirus 1 and 2 and enterovirus 90

while the smaller clade contained echovirus 11 and 13,

enterovirus B, and coxsackievirus A16 (Fig. 3). The large

cluster could be further divided into two groups with one

group comprising poliovirus 2 only and the other group

containing enterovirus 90 and poliovirus 1. The most

prevalent enterovirus was poliovirus 2 which was isolated

from 11 out of 15 samples. Two samples were positive for

enterovirus 90 and three were positive for poliovirus 1.

3.4. Concentrations of thermotolerant coliforms in VeniceLagoon and beach samples and their relationship to levels ofHAV and enterovirus

Thermotolerant coliform levels ranged from 5000 to

540,000 CFP/L (colony forming particles per liter) in the Venice

canals and from o1 to 155 CFP/L at Lido beach (Table 3). There

was a statistically significant correlation (R2¼ 0.62, p ¼ 0.0002)

between thermotolerant coliform densities and HAV levels, but

the relationship between thermotolerant coliforms and enter-

ovirus was not significant (R2¼ 0.08, p ¼ 0.2572) (Fig. 4).

4. Discussion

Indicators for assessing water quality have been a subject of

some controversy for over 50 years. Waterborne marine

illnesses are most often associated with viruses rather than

bacteria (Griffin et al., 2003). However, current bathing water

quality requirements in European Union countries are based

on levels of fecal bacteria indicators (thermotolerant coli-

forms) rather than virus. Thermotolerant coliform indicators

have been shown to die off more quickly in seawater than

many viruses, and therefore may not be found in contami-

nated water where viruses can still persist (Fattal et al., 1983).

Therefore, in order to assess the human health risk asso-

ARTICLE IN PRESS

Fig. 4 – Regression analysis of thermotolerant coliform densities as a function of (A) HAV and (B) enterovirus concentrations in

seawater, collected from the Venice Lagoon canals and Lido beach in Venice, Italy. Non-detectable (ND) levels of bacteria and

virus were assigned the value of one-half of the limit of detection.

WAT E R R E S E A R C H 40 (2006) 2387– 2396 2393

ciated with exposure to sewage-contaminated waters of the

Venice Lagoon, we used real time RT-PCR to measure levels of

HAV and enteroviruses directly, and to better define the

relationship between these viruses and thermotolerant coli-

forms.

To date, there have been few studies performed to evaluate

the viral water quality of beaches and coastal waters along

the Adriatic coastline of Italy. Muscillo et al. (1999) used RT-

PCR to detect poliovirus 3 in estuarine waters of the Foglia

River and along the beaches of Pesaro along the Adriatic coast

of Italy. Another study by a different group of researchers in

the same area of the Adriatic coast, detected the presence of

enterovirus by cell culture in 32% of 144 samples (Pianetti et

al., 2000). These authors concluded that viral pollution

originating from regional public and resort wastewater

disposal systems could negatively impact regional beach

water quality. In another survey of water quality in the

Adriatic Sea near Fano, Italy, Muscillo et al. (2001) used RT-PCR

to identify reoviruses in 30% of 72 seawater samples; however,

no enteroviruses were detected. The fact that Muscillo et al.

(2001) did not detect enteroviruses in their study while we did

in most of our samples may have been due to better quality

water samples in their study or to the generally superior

detection sensitivity of real-time RT-PCR versus conventional

RT-PCR.

Although the development of real-time RT-PCR methods

now makes it possible to quantitate levels of human viruses

in other impacted coastal waters like the Venice Lagoon, such

molecular detection techniques cannot distinguish between

infectious and non-infectious particles. On the other hand,

since HAV detection and quantification by conventional cell

culture assay is often difficult and time-consuming, there is

little data that has been published to date on HAV levels in

impacted marine waters. Moreover, despite its limitations in

recognizing infectivity, real-time RT-PCR can still be valuable

as an indicator of recent viral contamination (Gantzer et al.,

1999). Using real-time PCR, Brooks et al. (2005) detected HAV

in all eight samples taken during rain events from either the

mouth of the Tijuana River (near the US–Mexico border) or the

nearby surfzone at Imperial Beach, CA, at levels ranging from

90 to 3523 and 347 to 2656 copies/L, respectively. These

relatively high levels of HAV measured during wet weather

were attributed to the inadequate sewage collection infra-

structure in the region of Tijuana, Mexico.

In the present study, both HAV and enteroviruses were

detected in 78% of the canal samples analyzed, with levels

(uncorrected for recovery efficiency) ranging from 75 to 730

and 3 to 1614 copies/L, respectively (Table 3). It is important to

note here that Venice hosts as many as 14 million tourists per

year from all parts of the world. This, coupled with the

inadequate sewage infrastructure in the city of Venice,

suggest that the relatively high levels of these viruses that

we measured may not be characteristic of other urban coastal

waters in Europe. On the other hand, the occurrence of

enteroviruses in European coastal marine waters has been

previously reported for the Mediterranean Sea off Italy

(Muscillo et al., 1999, 1994; Pianetti et al., 2000), along the

beaches of southwest Greece (Vantarakis and Papapetropou-

lou, 1998), and the coastal waters of Northern Ireland (Hughes

et al., 1992).

Extensive studies of the Florida Keys have shown wide-

spread bacterial and viral contamination of nearshore surface

waters often associated with septic systems for sewage

disposal. In a total of 17 canal sites and 2 nearshore water

sites, 79% of the samples sites were positive for enterovirus,

63% positive for HAV, and 11% positive for Norwalk viruses

when samples were assayed by RT-PCR (Griffin et al., 1999). In

southern California coastal waters near the US–Mexico

border, Jiang et al. (2001) found 33% (4 of 12) of marine

samples were positive for adenoviruses as determined by

nested PCR. Interestingly, at these marine sites located

outside of river discharge points, Jiang et al. (2001) noted that

bacteria indicators did not correlate with the presence of

viruses. In our study, concentrations of HAV at the beach on

Lido island were always below the level of detection (Table 3).

However, enteroviruses were detected in all Lido samples at

relatively low levels (uncorrected for recovery efficiency)

ranging from 2 to 71 copies/L (Table 3).

Our seawater samples were processed following a pub-

lished virus concentration method (using negatively charged

filters) for enteroviruses and noroviruses where the recovery

was reported to be 60–70% (Katayama et al., 2002). However, in

ARTICLE IN PRESS

WAT E R R E S E A R C H 4 0 ( 2 0 0 6 ) 2 3 8 7 – 2 3 9 62394

a more recent study by Fuhrman et al. (2005), using the

Katayama et al. (2002) protocol but modified for subsequent

RNA extraction of the concentrate, recovery efficiencies were

only in the range of 12.3–22.6% for seeded poliovirus in

seawater. In our study, the recovery efficiency of poliovirus

seeded into natural seawater was found to be relatively high

(71%), more comparable to the original results of Katayama et

al. (2002) (Table 2). On the other hand, although the recovery

efficiency of HAV from seawater by the Katayama et al. (2002)

protocol has not been previously reported, we found it to be

rather low (12%) (Table 2). Since the levels of HAV and

enteroviruses in the Venice canals were more or less

comparable, this suggests that the poorer HAV recovery

efficiency might explain why enteroviruses were sometimes

detected at Lido beach, while HAV was not.

The sequence data from all the HAV clones isolated for both

years 2003 and 2004 showed 100% similarity indicating the

prevalence of predominantly a single isolate of HAV in the

sampled region of the Venice Lagoon. Ticehurst et al. (1988)

reported that different human HAV strains of diverse

geographic origin were remarkably closely related. This

isolate was almost identical (98%) to an isolate from Mexico

(accession #AY441441) and only slightly less similar to

isolates from southern Italy (97%; accession #AJ505803),

Argentina (96%; accession #AF452067), and Japan (95%;

accession #AB020569).

A BLAST search of enterovirus sequences identified seven

different enterovirus types. This suggests that although the

primer sequences in the 50-UTR are highly conserved among

enteroviruses, the sequences flanking these primers are

variable enough to distinguish different enterovirus types.

However, the short length of the amplicon (151 bp) did not

allow for discrimination between enteroviral strains (e.g.,

wild type and vaccine strain polioviruses). Among the clones

sequenced, poliovirus 2 was most prevalent followed by

poliovirus 1 and enterovirus 90. A neighbor-joining tree

grouped the enterovirus isolates into two major clades. One

clade contained poliovirus 1 and 2, and enterovirus 90

whereas the second clade contained coxsackievirus A16,

enterovirus B, echovirus 11, and echovirus 13. This is in

general agreement with previously published enterovirus

phylogeny (Muir et al., 1998).

Since we found a predominant type of HAV and enterovirus

(poliovirus 2, the same type as our positive control), one

might argue that this resulted from a contamination event in

the laboratory. However, this is unlikely for several reasons.

First, negative controls run in parallel with positive samples

were consistently negative by both PCR and sequencing. In

addition, four of the water samples positive for poliovirus 2

also contained at least one other type of enterovirus. Finally,

levels of HAV and poliovirus 2 were highly variable among the

water samples.

Despite our detection of low levels of enteroviruses in all

the Lido beach samples, it is important to note here that in

these same samples, thermotolerant coliform levels ranged

from 0 to 155 CFP/L at Lido beach and never exceeded the

criteria under Italian national law (Presidential Decree no. 470

of 1982 which acknowledges European Council Directive 76/

160/EEC on Bathing Water Quality) of an upper limit

(imperative value) for thermotolerant coliforms of 2000 CFP/

100 mL. Indeed, only a single sample taken on 25 May 2005

exceeded the guideline value of 100 CFP/100 mL. Moreover,

while our analysis showed there was a statistically significant

(p ¼ 0.0002) relationship between densities of thermotolerant

coliform bacteria and HAV levels for all the samples pooled

(Fig. 4), there was no significant relationship (p ¼ 0.2572)

evident between thermotolerant coliform indicator densities

and enterovirus levels. This latter result supports the need for

the development of both rapid and sensitive methods to

quantitate human pathogens directly rather than relying on

the conventional bacterial indicators to assess human health

risk in recreational marine waters.

We attempted to interpret enteroviral levels in terms of a

quantitative risk assessment for swimming at Lido beach, by

relating the PCR-quantified viral densities to infectivity.

Donaldson et al. (2002) concluded from data of a side-by-side

comparison of cell culture and real-time RT-PCR for enter-

oviruses, that 55 viral particles in a sample equates to one

infectious particle. Using this infectivity relationship for

enterovirus would equate to 0.04–1.3 infectious particles/L

for Lido beach. Assuming an incidental ingestion of 100 mL of

seawater during swimming, then the risk of infection may

then be calculated by using the beta-Poisson model (Regli

et al., 1991):

Pi ¼ 1� 1þmVb

� ��a(3)

where Pi is the probability of infection resulting from

ingestion of a single volume V of water containing an average

of m organisms per unit volume, and a and b are model

parameters that characterize the dose-response curve ex-

posure. Using the best-fit model parameters of a ¼ 0.409 and

b ¼ 0.788 for poliovirus III (Regli et al., 1991) as being most

conservatively representative of the infectivity of entero-

viruses, the daily risk of enteroviral infection for exposure at

Lido beach can be calculated to range from 1.9�10�3 to

6.1�10�2 with the lowest risk detectable by our method at

1.1�10�3. It is noted here that the risk estimates do not take

into account our recovery efficiency of enteroviruses, but

since the efficiency was relatively high (71%) (Table 2), the risk

outcomes would not be significantly changed by such a

correction. A sensitivity analysis of the risk outcomes using

the beta-Poisson model showed that over a range of 71 order

of magnitude from the base values, the model was nearly

equally sensitive to changes in each model parameter

(though in an inverse way for b).

Since the risk of symptomatic disease may range from 1

percent for poliovirus to more than 75 percent for some of the

coxsackie B viruses (Cherry, 1981), then the daily risk of

disease may be assumed to be no higher than about 4.5�10�2

even at the highest enterovirus level we measured. Such a

conclusion suggests that bathers at Venice’s Lido beach are at

or below the disease risk (5% for gastroenteritis) that is

deemed acceptable by complying with the standards of the

European Directive (Commission of the European Commu-

nities, 2002). It should be noted here, however, that enter-

oviruses can cause a variety of other and more serious disease

symptoms besides (or in addition to) gastroenteritis including

poliomyelitis, aseptic meningitis and myocarditis, but these

ARTICLE IN PRESS

WAT E R R E S E A R C H 40 (2006) 2387– 2396 2395

diseases are not subsumed within the risk outcomes of the

European Union Directive.

Real-time RT-PCR is relatively rapid with respect to current

water quality monitoring procedures, with an entire proces-

sing time of less than 12 h. In addition, this method has the

potential to offer greater sensitivity and quantitative ability

than any single method currently offers. With further

optimization of viral concentration procedures, the applic-

ability of this method to high throughput reproducible assays

could be developed for routine detection of human pathogens

in marine recreational waters impacted by sewage contam-

ination such as the Venice Lagoon.

5. Conclusions

Venice canal samples were often contaminated with high

levels of both HAV and enteroviral genomes, reflecting the

high degree of sewage contamination of these waters. At the

beach on Lido island, concentrations of HAV were always

below the level of detection, and enteroviruses (when

detected), were always present at relatively low levels.

The risk for enteroviral infection (calculated using the beta-

Poisson model) for recreational exposure from swimming at

Lido beach was in the range of 1.9�10�3–6.1�10�2, yielding a

disease risk at or below the level deemed acceptable by

European Guide standards.

There was a statistically significant correlation between

thermotolerant coliform densities and HAV levels, but not

between thermotolerant coliforms and enterovirus levels,

supporting the need for methods to quantitate human viruses

directly rather than relying on the conventional bacterial

indicators.

Acknowledgements

We thank the San Diego State University Research Foundation

and the Office of International Programs of San Diego State

University for financial support. We also thank the Southwest

Center for Environmental Research Policy (SCERP) for techni-

cal support and Walter Hayhow for technical assistance in the

laboratory.

R E F E R E N C E S

Aimo, E., Della Sala, S., Stradella, S.,Tagliapietra, D., Vazzoler, M.,1999. L’inquinamento dei rii: La diffusione degli inquinanti neicanali interni di Venezia. In: Venezia; la citta dei rii. UNESCO-Insula S.p.a., Venezia.

Altschul, S.F., Madden, T.L., Schaffer, A.A., Zhang, J., Zhang, Z.,Miller, W., Lipman, D.J., 1997. Gapped BLAST and PSI-BLAST: anew generation of protein database search programs. NucleicAcids Res. 25, 3389–3402.

American Public Health Association (APHA), 1992. StandardMethods for the Examination of Water and Wastewater.American Public Health Association, Washington, DC.

Bernstein, A.G., Cecconi, G., 1996. The impact of the Venice tidalbarriers on water quality. In: Proceedings of the InternationalConference on Barrages, Cardiff, UK, September 10–13, 1996.

Blanchard, D.C., 1989. The ejection of drops from the sea and theirenrichment with bacteria and other materials: a review.Estuaries 12 (3), 127–137.

Brooks, H.A., Gersberg, R.M., Dhar, A.K., 2005. Detection andquantification of hepatitis A virus in seawater via real-time RT-PCR. J. Virol. Methods 127 (2), 109–118.

Cherry, J.D., 1981. Non-polio enteroviruses: coxsackieviruses,echoviruses and enteroviruses. In: Feigin, R.D., Cherry, J.D.(Eds.), Textbook of Pediatric Infectious Diseases. Saunders,Philadelphia, Pennsylvania.

Commission of the European Communities, 2002. Proposal for aDirective of the European Parliament and of the CouncilConcerning the Quality of Bathing Water, Brussels, Belgium.

Deffernez, C., Wunderli, W., Thomas, Y., Yerly, S., Perrin, L., Kaiser,L., 2004. Amplicon sequencing and improved detection ofhuman rhinovirus in respiratory samples. J. Clin. Microbiol. 42(7), 3212–3218.

Donaldson, K.A., Griffin, D.W., Paul, J.H., 2002. Detection, quanti-tation and identification of enteroviruses from surface watersand sponge tissue from the Florida Keys using real-time RT-PCR. Water Res. 36, 2505–2514.

Fattal, B., Vasl, R.J., Katzenelson, E., Shuval, H.I., 1983. Survival ofbacteria indicator organisms and enteric viruses in theMediterranean coastal waters off Tel-Aviv. Water Res. 17,397–402.

Fuhrman, J.A., Liang, X., Noble, R.T., 2005. Rapid detection ofenteroviruses in small volumes of natural waters by real-timequantitative reverse transcriptase PCR. Appl. Environ. Micro-biol. 71, 4523–4530.

Gantzer, C., Senouci, S., Maul, A., Levi, Y., Schwartzbrod, L., 1999.Enterovirus detection from wastewater by RT-PCR and cellculture. Water Science and Technology 40, 105–109.

Griffin, D.W., Donaldson, K.A., Paul, J.H., Rose, J.B., 2003. Patho-genic human viruses in coastal waters. Clinical MicrobiologyReviews 16, 129–143.

Griffin, D.W., Gibson III, C.J., Lipp, E.K., Riley, K., Paul III, J.H., Rose,J.B., 1999. Detection of viral pathogens by reverse transcriptasePCR and of microbial indicators by standard methods in thecanals of the Florida keys. Appl. Environ. Microbiol. 65,4118–4125.

Haramoto, E., Katayama, H., Oguma, K., Ohgaki, S., 2005.Application of cation-coated filter method to detection ofnoroviruses, enteroviruses, adenoviruses, and torque tenoviruses in the Tamagawa River in Japan. Appl. Environ.Microbiol. 71 (5), 2403–2411.

Hughes, M.S., Coyle, P.V., Connolly, J.H., 1992. Enteroviruses inrecreational waters of Northern Ireland. Epidemiology andInfection 108 (3), 529–536.

Jiang, S., Noble, R., Chu, W.P., 2001. Human adenoviruses andcoliphages in urban runoff-impacted coastal waters ofSouthern California. Applied and Environmental Microbiology67, 179–184.

Johnston, P.A., MacGarvin, M., Stringer, R.L., Troendle, S., Swin-dlehurst, R.J., 1993. Sewage: towards realistic environmentalprotection. Water Sci. Technol. 27, 481–491.

Katayama, H., Shimasaki, A., Ohgaki, S., 2002. Development of aVirus Concentration method and its application to detectionof enterovirus and Norwalk virus from coastal seawater. Appl.Environ. Microbiol. 68, 1033–1039.

Kumar, S., Tamura, K., Nei, M., 2004. MEGA3: Integrated softwarefor Molecular Evolutionary Genetics Analysis and sequencealignment. Briefings Bioinform. 5, 50–163.

Leutenegger, C.M., Mislin, C.N., Sigris, B., Ehrengruber, M.U.,Hofmann-Lehmann, R., Lutz, H., 1999. Quantitative real-timePCR for the measurement of feline cytokine mRNA. Vet.Immunol. Immunopathol. 71, 291–305.

Mohamed, N., Elfaitouri, A., Fohlman, J., Friman, G., Blomberg, J.,2004. A sensitive and quantitative single-tube real-time

ARTICLE IN PRESS

WAT E R R E S E A R C H 4 0 ( 2 0 0 6 ) 2 3 8 7 – 2 3 9 62396

reverse transcriptase-PCR for detection of enteroviral RNA. J.Clin. Virol. 30 (2), 150–156.

Muir, P., Kammerer, U., Korn, K., Mulders, M.N., Poyry, T.,Weissbrich, B., Kandolf, R., Cleator, G.M., van Loon, A.M.,1998. Molecular typing of enteroviruses: current status andfuture requirements. The European Union Concerted Actionon Virus Meningitis and Encephalitis. Clin. Microbiol. Rev. 11(1), 202–227.

Muscillo, M., Aulicino, F.A., Patti, A.M., Orsini, P., Volterra, L., Farra,G.M., 1994. Molecular techniques for the identification ofenteric viruses in marine waters. Water Research 28, 1–7.

Muscillo, M., La Rosa, G., Carducci, A., Cantiani, L., Marianelli, C.,1999. Molecular and biological characterization of poliovirus 3strains isolated in Adriatic seawater samples. Water Res. 33(14), 3204–3212.

Muscillo, M., La Rosa, G., Marianelli, C., Zaniratti, S., Capobianchi,R., Cantiani, L., Carducci, A., 2001. A new RT-PCR method forthe identification of reoviruses in seawater samples. WaterRes. 35 (2), 548–556.

Orlob, G.T., Bale, A.E., Rajbhandri, H., Malagoli, M., 1991. Modelingeffects of tidal barrier closure of Venice Lagoon. Water Sci.Technol. 24, 149–155.

Pavoni, B., Sfriso, A., Donazzolo, R., Orio, A.A., 1990. Influence of

waste waters from the city of Venice and the hinterland on the

eutrophication of the lagoon. Sci. Total Environ. 96, 235–252.

Pianetti, A., Baffone, W., Citterio, B., Casaroli, A., Bruscolini, F.,

Salvaggio, L., 2000. Presence of enteroviruses and reoviruses in

the waters of the Italian coast of the Adriatic sea. Epidemiol.

Infect. 125, 455–462.

Regli, S., Rose, J.B., Haas, C.N., Gerba, C.P., 1991. Modeling the risk

from Giardia and viruses in drinking water. J. Am. Water

Works Assoc. 83, 76–84.

Ticehurst, J.C., Cohen, J.I., Purcell, R.H., 1988. Analysis of

molecular sequences demonstrates that hepatitis A virus

(HAV) is a unique picornavirus. In: Zuckerman, A.J. (Ed.), Viral

Hepatitis and Liver Disease. Alan R. Liss, New York, pp. 33–35.

Vantarakis, A.C., Papapetropoulou, M., 1998. Detection of enter-

oviruses and adenoviruses in coastal waters of SW Greece by

nested polymerase chain reaction. Water Research 32,

2365–2372.

Vazzoler, M., Stradella, S., 1999. La situazione sanitaria dei canali

interni di Venezia. In: Relazione finale sul progetto UNESCO

‘‘Canali interni di Venezia’’ anno 1998–1999, Venezia, Italy.